Nagios Support Forum

Hi All,

I met a problem on distributed Nagios, deployed by mod_gearman (version:3). I built a eventhandler for a certain service (check windows service), if this windows service stopped, this eventhandler will start it. HOwever, this eventhandler only work at Nagios Master ( with Mod_gearman module installed), not work on those workers. I am wondering I can seek some support/suggestions.

THanks.

Below is my configuration on /etc/mod_gearman/module.cfg at Nagios Master

#Mod-Gearman NEB Module Config
eventhandler=no
notifications=no
services=yes
hosts=yes
hostgroups=XXX_Host_Group,YYY_Host_Group,ZZZ_Host_Group
servicegroups=XXX_Service_Group,YYY_Service_Group,ZZZ_Service_Group

do_hostchecks=yes
route_eventhandler_like_checks=yes
encryption=yes
mum length is 32 characters.
key=******
use_uniq_jobs=on

This is configuration on /etc/mod_gearman/worker.cfg at Nagios slave

#Mod-Gearman NEB Module Config
eventhandler=yes
notifications=no
services=no
hosts=no
hostgroups=XXX_Host_Group
servicegroups=XXX_Service_Group

Change this in the /etc/mod_gearman/module.cfg:
to:
A copy of the handler should exist on the mod gearman workers - it usually needs to be under /usr/local/nagios/libexec or wherever the plugins are stored. Restart the modgearman service after making the change and verifying a copy of the event handler is on the gearman worker.

Hi, Thanks for your reply.

I've tried

1. change the eventhandler = yes at Nagios master, and stop nagios--> restart mod gearman --> start nagios. and perform the nagios and work.service restart at Nagios Slave
2. stop the test service at test PC, but no luck.

at Nagios Master, /usr/local/nagios/var/nagios.log. it shows SERVICE ALERT: XXXX;Check_Service;CRITICAL;HARD;3;CRITICAL: spooler=stopped (auto), delayed (). There is nothing shown at Nagios Slave. I checked, that script in the correct path at Nagios Slave. I ran this script manually, it can restart the target service. So would you please provide some inputs about this. THanks very much.

To get gearman to handle events you would also need to edit your /usr/local/nagios/etc/nagios.cfg and remove the eventhandler=no option from the mod_gearman broker line and restart the nagios service:
If you do that you would need to make sure all the event handler plugins exist on the workers as well.

Make sure you do this on a test system first to make sure it doesn't impact your monitoring/notifications.

Hi, Thanks very much, I've followed your proposal and got the successful result. Thanks Both again. These inputs are very helpful

Great!! .... ticket locked!!